Unless any of these recommendations are accompanied by published test data on what happens under load, I wouldn’t place too much weight on this issue.
I’ll stick with what Shimano, Sapim and Magura recommend, given they make and test wheels, spokes and disc brakes or all three, even if they don’t publish their test results. Who exactly is recommending the opposite approach, except for Chris King.
There are arguments based on what happens under load, but these work both ways. One type of argument is that the crossing point should be pulled in under load to prevent fouling – but that means lacing one way on the disc side (prevent fouling on the caliper under braking) and the other way on the rear drive side (prevent fouling on the derailleur cage when pedalling in first gear). And spoke/caliper fouling may not be an issue with your setup. Another argument is that the outside spokes should be the ones loaded under braking as those spokes are better supported by the flange. And so on
This is nonsensical. All your arguments are in line with the Shimano/sapim/Magura pattern and logic! So what if you have to lace differently on each side on the rear? Look at the diagram, you have the outside spokes supported by the flange and being tensioned under braking.
Replace the nipples though.
Why would you replace the nipples from a new wheel? If you make a mistake and have to relace a wheel would you throw the nipples away?